GMStool

GMStool selects optimal single nucleotide polymorphism (SNP) marker sets from genome-wide association study (GWAS) results to improve genomic prediction of continuous (quantitative) phenotypes.

Key Features:

• GWAS-Based Approach: Utilizes genome-wide association studies (GWAS) to identify candidate genetic markers.
• SNP Marker Focus: Targets single nucleotide polymorphism (SNP) markers for phenotype estimation.
• Heuristic Search Algorithm: Employs heuristic search techniques to explore combinations of markers.
• Integration of Statistical and Machine Learning Models: Integrates traditional statistical models and machine/deep-learning algorithms for genomic prediction.
• Optimal Marker Set Selection: Selects reduced marker sets intended to maximize prediction accuracy compared with full marker sets or top GWAS markers.

Scientific Applications:

• Phenotype Prediction: Predicts quantitative (continuous) phenotypes from genomic data.
• Comparative Performance: Applied to improve predictive performance relative to methods using full marker sets or top GWAS-identified markers.

Methodology:

Uses GWAS to identify candidate markers, then systematically searches marker combinations using heuristic search combined with statistical models and machine/deep-learning algorithms to select marker sets that maximize prediction accuracy.